Generative AI faces mounting pressure over costs and financing
Generative AI businesses face growing financial pressure as intense competition, falling prices and expensive model development constrain profitability. Similar training methods and datasets make lasting competitive advantages difficult to establish. The Claude 5.5 model family reportedly cost $10 billion to develop, while inference expenses add further pressure. Reliability problems, weak instruction-following and uncertain returns on investment continue to complicate adoption.
Market declines are adding to concerns about the sector’s economics. Nvidia is down 29% over the last six months, while Oracle has fallen 35% and CoreWeave 54%. Anthropic’s perceived valuation has dropped 25% in a matter of weeks. Companies rehiring humans after ambitious AI replacement claims, including Klarna, also suggest that practical performance has not consistently matched expectations.
A potential retreat from Nvidia’s financing commitments could put additional strain on infrastructure providers and model developers. Nscale is described as financially strained after overcommitting, while a more cautious approach under Nvidia’s interim leadership could reduce support for dependent businesses. Tighter lending conditions could further raise funding costs, potentially threatening expansion plans and Anthropic’s IPO prospects. The central risk is whether these companies can operate sustainably without continued financial backing.